Ethical Analysis: Predictive Risk Models at Call Screening for Allegheny County.
Dare, Tim. Gambrill, Eileen.
Published: April 2017
Allegheny County Pennsylvania Department of Human Services
The Human Services Building One Smithfield Street
Pittsburgh, PA 15222
This report comments on two linked papers produced by Rhema Vaithianathan, Nan Jiang, Tim Maloney and Emily Putnam-Hornstein as part of the development of a predictive risk modeling tool to improve child protection decisions being made by the Allegheny County Department of Human Service. In our assessment, subject to the recommendations in this report, the implementation of the AFST is ethically appropriate. Indeed, we believe that there are significant ethical issues in not using the most accurate risk prediction measure. Instruments that are more accurate will result in fewer false positives and false negatives, thus reducing stigmatization (false positives) and more lost opportunities to protect children. It is hard to conceive of an ethical argument against use of the most accurate predictive instrument. As we have emphasized throughout, decisions are being made right now. It is not a matter ofmaking or not making related decisions. The decisions involved are complex ones made in a context of inevitable uncertainty that contributes to inevitable error. Research on decision-making in the helping professions highlights the play of biases and fallacies. Confirmation biases are common in which we seek information that corresponds to our preferred view (e.g., there is no abuse) and fail to seek evidence that contradicts preferred views. Errors of omission (failing to act) are viewed as less harmful than errors of commission (acting - for example, removing a child from the care of her family). The question is, how can we make the fewest errors in our efforts toprotect children and families? AFST seems an ethical and potentially important contribution to that effort. (Author abstract modified)
PREDICTIVE ANALYTICS; CHILD PROTECTIVE SERVICES; DECISION MAKING; RISK ASSESSMENT; RISK FACTORS; PREDICTOR VARIABLES; MODELS; DECISION MAKING; SERVICE DELIVERY; PROGRAM IMPLEMENTATION; PROGRAM PLANNING; RESEARCH METHODOLOGY; ETHICS; PENNSYLVANIA